随着认知语言学基于使用的语言学和语言习得理论的发展，语言中各类由两个及以上词项构成的语块获得越来越多的关注。从心理加工层面来看，一方面，语块的突出特点是整体加工，在句子中不需要经过逐词提取和组装，而是直接整体提取，这一特点使其能为在线语言加工节省认知资源，降低认知负荷。但另一方面，语块的整体加工也可能削弱其内部成分词的表征和加工。语块整体加工的双重特性如何影响汉语二语特殊句式的习得，还有待进一步考察。

在此背景下，本研究采取被试间的实验设计和前测-后测的实验范式，考察“语块呈现”和“非语块呈现”两种结构呈现方式对汉语二语“把”字句不同维度知识、技能学习效果的影响及其原因。两种呈现方式的区别主要在于，“语块呈现”条件下以空格这一视觉线索对学习句进行切分，使得切分后的句子包含介宾语块和动补语块这两类语块，如“男孩 把杯子 打碎 了”；“非语块呈现”条件下则在句中所有单词间插入空格，切分后的句子不包含语块，如“男孩 把 杯子 打 碎 了”。各维度知识、技能及其对应的学习效果测试任务包括：语义知识-二选一式句图匹配任务、结构知识-句子合法性判断任务，阅读技能-自定步速句子阅读任务、产出技能-口头组词成句任务。

本研究的实验结果显示：（1）在四项学习效果测试任务的前、后测比较中，两组被试后测反应时相比于前测显著下降，正确率显著上升，体现了学习阶段的输入带来的学习效应；（2）在四项学习效果测试任务的后测组间比较中，两组被试间未出现反应时或正确率的显著差异；（3）在句子合法性判断任务组间后测的细化比较中，两组被试在正确句、“把”字遗漏句、动词遗漏句三类测试句的判断反应时上表现出不同的反应模式。语块组被试对“把”字遗漏句的判断反应时显著慢于正确句和动词遗漏句，非语块组被试则未表现出测试句类型间的显著差异。针对上述实验结果及其与以往研究结果的比较，本研究得出的主要结论如下：（1）从语块呈现和非语块呈现对语义知识和结构知识学习效果的影响看，呈现方式对语义知识的影响可能受到输入通道的调节作用，相比于视觉输入通道，学习者对听觉输入通道中的语流切分线索更为敏感。呈现方式对结构知识的影响主要表现为语块整体加工对语块内部成分词的削弱，导致学习者对句法结构完整性的判断出现偏差，且这一削弱受到成分词词性和语义特征的调节；（2）从语块呈现和非语块呈现对阅读技能和产出技能的影响看，语块整体加工在阅读任务中的加工优势可能会受到自定步速阅读任务中分词呈现特征的干扰，在“把”字句的口语产出中可能带来的负面作用则在语言产出本身心理加工过程的影响下被淹没；（3）总体而言，实验中采取的空格这一视觉切分线索对学习者在句子加工中注意资源分配方式的影响可能有限，且受到学习者先前知识经验的调节作用。与下划线、粗体、斜体等文本强化类似，语块间的空格作为一种隐性、间接的注意引导方式很难单独对学习效果产生明显影响。要发挥视觉线索的作用，可能需要增加干预频次、干预时长，或将视觉线索与其他外显教法（如讲解）结合进行干预。此外还需考虑到，若学习者同时具备自上而下的知识线索和自下而上的感知觉线索，前者的影响将更大，且可能淹没甚至压制后者的作用。

本研究对结构呈现方式如何影响汉语二语“把”字句不同维度知识、技能的学习进行了细化考察，将语块在认知加工上对二语习得的积极作用延伸至二语特殊句式学习中，为语块整体加工和结构呈现方式在二语特殊句式教学中的应用提供了一定实证证据的支撑。

The usage-based approach to second language acquisition has led to increased attention on chunks composed of multiple lexical item. The main feature of chunks is holistic processing, allowing for direct extraction as a whole and reducing cognitive load during online language processing. However, the holistic processing may also weaken of the representation and processing of its constituent words. Further investigation is needed to understand how the dual characteristics of chunk’s holistic processing impact second language acquisition of Chinese special sentence patterns.

This study adopts a between-subject and pretest-posttest design to investigate the influence of two structure presentation modes, i.e., "chunk presentation" and "non-chunk presentation", on the learning outcomes of various dimensions of knowledge and skills of ba sentence in Chinese. The difference between the two presentation modes primarily lies in the following aspects. In the "chunk presentation" condition, the learning sentence is segmented with visual cues of space between chunks, resulting in two types of chunks in the sentence, the preposition-object chunk and the verb-complement chunk. For example, "男孩 把杯子 打碎 了". On the contrary, the "non-chunk presentation" condition inserts spaces between all the words in the sentence, leading to a segmented sentence without any chunks. For example, "男孩 把 杯子 打 碎 了". The testing of learning outcomes includes two alternative sentence picture matching task for semantic knowledge, grammatical judgment task for structural knowledge, self-paced sentence reading task for reading skill, and verbal grouping words into sentence task for production skill.

Mixed-effects modeling revealed that (a) both groups exhibited a significant decrease in reaction time and increase in accuracy rate from pretests to posttests, indicating a learning effect resulting from the input of the learning stage; (b) there was no significant difference in reaction time or accuracy rate between the two groups in the posttests; (c) in the grammatical judgement task, the two groups displayed different patterns of reaction time for correct sentence, ba missing sentence, and verb missing sentence in the posttests. The chunk group showed significantly slower reaction time for ba missing sentence compared to correct sentence and verb missing sentence, while the non-chunk group did not exhibit significant difference among these three types of testing sentence.

The results are discussed from the following perspectives: (a) with regard to the impact of chunk and non-chunk presentation on learning semantic and structural knowledge, the influence on semantic knowledge may be modulated by input modality. Compared with visual input, learners tend to demonstrate greater sensitivity to segmentation cues in auditory input. The effect of structure presentation on structural knowledge is primarily evident in that the holistic processing of chunks weakens the constituent words within chunks, leading to learners' deviation in judging the grammaticality of syntactic structure, and this weakening effect is constrained by the word class and semantic features of the constituent words; (b) in terms of the effect of chunk and non-chunk presentation on reading and speaking skills, the holistic processing advantages of chunk in reading tasks may be disrupted by the word-by-word nature of self-paced reading task. As for the verbal production of ba sentence, the potential negative effects resulting from the weakened constituent words tend to be overshadowed by the influence of the language production mechanism itself; (c) on the whole, the visual segmentation cue used in the experiment has limited influence on the allocation of attention resources in learners’ sentence processing, and is likely to be modulated by the prior knowledge of the learners. Similar to textual enhancement such as underlining, bold, italics, etc., which serve as an implicit and indirect way to guide attention, it’s difficult for the spacing between chunks alone to affect learning. To make visual cues effective, an increase in intervention frequency and duration, or the combination of visual cues with other explicit teaching methods (such as explanation) might be required. Besides, it needs to be taken into consideration that, if learners have access to both top-down knowledge and bottom-up sensory and perceptual cues, the former will have a greater impact and may submerge or even suppress the effect of the latter.

 The current study closely examined the impact of structure presentation on the learning of different dimensions of knowledge and skills of ba sentence, further highlighting the beneficial role of holistic processing of chunks in second language acquisition of Chinese special sentence patterns. It also offers empirical support for the application of chunk processing and structure presentation in second language teaching of Chinese special sentence pattern.